Home | History | Annotate | Line # | Download | only in kern
kern_clock.c revision 1.94.4.2
      1 /*	$NetBSD: kern_clock.c,v 1.94.4.2 2006/12/30 20:50:05 yamt Exp $	*/
      2 
      3 /*-
      4  * Copyright (c) 2000, 2004 The NetBSD Foundation, Inc.
      5  * All rights reserved.
      6  *
      7  * This code is derived from software contributed to The NetBSD Foundation
      8  * by Jason R. Thorpe of the Numerical Aerospace Simulation Facility,
      9  * NASA Ames Research Center.
     10  * This code is derived from software contributed to The NetBSD Foundation
     11  * by Charles M. Hannum.
     12  *
     13  * Redistribution and use in source and binary forms, with or without
     14  * modification, are permitted provided that the following conditions
     15  * are met:
     16  * 1. Redistributions of source code must retain the above copyright
     17  *    notice, this list of conditions and the following disclaimer.
     18  * 2. Redistributions in binary form must reproduce the above copyright
     19  *    notice, this list of conditions and the following disclaimer in the
     20  *    documentation and/or other materials provided with the distribution.
     21  * 3. All advertising materials mentioning features or use of this software
     22  *    must display the following acknowledgement:
     23  *	This product includes software developed by the NetBSD
     24  *	Foundation, Inc. and its contributors.
     25  * 4. Neither the name of The NetBSD Foundation nor the names of its
     26  *    contributors may be used to endorse or promote products derived
     27  *    from this software without specific prior written permission.
     28  *
     29  * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
     30  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
     31  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
     32  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
     33  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
     34  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
     35  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
     36  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
     37  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
     38  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
     39  * POSSIBILITY OF SUCH DAMAGE.
     40  */
     41 
     42 /*-
     43  * Copyright (c) 1982, 1986, 1991, 1993
     44  *	The Regents of the University of California.  All rights reserved.
     45  * (c) UNIX System Laboratories, Inc.
     46  * All or some portions of this file are derived from material licensed
     47  * to the University of California by American Telephone and Telegraph
     48  * Co. or Unix System Laboratories, Inc. and are reproduced herein with
     49  * the permission of UNIX System Laboratories, Inc.
     50  *
     51  * Redistribution and use in source and binary forms, with or without
     52  * modification, are permitted provided that the following conditions
     53  * are met:
     54  * 1. Redistributions of source code must retain the above copyright
     55  *    notice, this list of conditions and the following disclaimer.
     56  * 2. Redistributions in binary form must reproduce the above copyright
     57  *    notice, this list of conditions and the following disclaimer in the
     58  *    documentation and/or other materials provided with the distribution.
     59  * 3. Neither the name of the University nor the names of its contributors
     60  *    may be used to endorse or promote products derived from this software
     61  *    without specific prior written permission.
     62  *
     63  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
     64  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     65  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     66  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
     67  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     68  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     69  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     70  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     71  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     72  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     73  * SUCH DAMAGE.
     74  *
     75  *	@(#)kern_clock.c	8.5 (Berkeley) 1/21/94
     76  */
     77 
     78 #include <sys/cdefs.h>
     79 __KERNEL_RCSID(0, "$NetBSD: kern_clock.c,v 1.94.4.2 2006/12/30 20:50:05 yamt Exp $");
     80 
     81 #include "opt_ntp.h"
     82 #include "opt_multiprocessor.h"
     83 #include "opt_perfctrs.h"
     84 
     85 #include <sys/param.h>
     86 #include <sys/systm.h>
     87 #include <sys/callout.h>
     88 #include <sys/kernel.h>
     89 #include <sys/proc.h>
     90 #include <sys/resourcevar.h>
     91 #include <sys/signalvar.h>
     92 #include <sys/sysctl.h>
     93 #include <sys/timex.h>
     94 #include <sys/sched.h>
     95 #include <sys/time.h>
     96 #ifdef __HAVE_TIMECOUNTER
     97 #include <sys/timetc.h>
     98 #endif
     99 
    100 #include <machine/cpu.h>
    101 #ifdef __HAVE_GENERIC_SOFT_INTERRUPTS
    102 #include <machine/intr.h>
    103 #endif
    104 
    105 #ifdef GPROF
    106 #include <sys/gmon.h>
    107 #endif
    108 
    109 /*
    110  * Clock handling routines.
    111  *
    112  * This code is written to operate with two timers that run independently of
    113  * each other.  The main clock, running hz times per second, is used to keep
    114  * track of real time.  The second timer handles kernel and user profiling,
    115  * and does resource use estimation.  If the second timer is programmable,
    116  * it is randomized to avoid aliasing between the two clocks.  For example,
    117  * the randomization prevents an adversary from always giving up the CPU
    118  * just before its quantum expires.  Otherwise, it would never accumulate
    119  * CPU ticks.  The mean frequency of the second timer is stathz.
    120  *
    121  * If no second timer exists, stathz will be zero; in this case we drive
    122  * profiling and statistics off the main clock.  This WILL NOT be accurate;
    123  * do not do it unless absolutely necessary.
    124  *
    125  * The statistics clock may (or may not) be run at a higher rate while
    126  * profiling.  This profile clock runs at profhz.  We require that profhz
    127  * be an integral multiple of stathz.
    128  *
    129  * If the statistics clock is running fast, it must be divided by the ratio
    130  * profhz/stathz for statistics.  (For profiling, every tick counts.)
    131  */
    132 
    133 #ifndef __HAVE_TIMECOUNTER
    134 #ifdef NTP	/* NTP phase-locked loop in kernel */
    135 /*
    136  * Phase/frequency-lock loop (PLL/FLL) definitions
    137  *
    138  * The following variables are read and set by the ntp_adjtime() system
    139  * call.
    140  *
    141  * time_state shows the state of the system clock, with values defined
    142  * in the timex.h header file.
    143  *
    144  * time_status shows the status of the system clock, with bits defined
    145  * in the timex.h header file.
    146  *
    147  * time_offset is used by the PLL/FLL to adjust the system time in small
    148  * increments.
    149  *
    150  * time_constant determines the bandwidth or "stiffness" of the PLL.
    151  *
    152  * time_tolerance determines maximum frequency error or tolerance of the
    153  * CPU clock oscillator and is a property of the architecture; however,
    154  * in principle it could change as result of the presence of external
    155  * discipline signals, for instance.
    156  *
    157  * time_precision is usually equal to the kernel tick variable; however,
    158  * in cases where a precision clock counter or external clock is
    159  * available, the resolution can be much less than this and depend on
    160  * whether the external clock is working or not.
    161  *
    162  * time_maxerror is initialized by a ntp_adjtime() call and increased by
    163  * the kernel once each second to reflect the maximum error bound
    164  * growth.
    165  *
    166  * time_esterror is set and read by the ntp_adjtime() call, but
    167  * otherwise not used by the kernel.
    168  */
    169 int time_state = TIME_OK;	/* clock state */
    170 int time_status = STA_UNSYNC;	/* clock status bits */
    171 long time_offset = 0;		/* time offset (us) */
    172 long time_constant = 0;		/* pll time constant */
    173 long time_tolerance = MAXFREQ;	/* frequency tolerance (scaled ppm) */
    174 long time_precision = 1;	/* clock precision (us) */
    175 long time_maxerror = MAXPHASE;	/* maximum error (us) */
    176 long time_esterror = MAXPHASE;	/* estimated error (us) */
    177 
    178 /*
    179  * The following variables establish the state of the PLL/FLL and the
    180  * residual time and frequency offset of the local clock. The scale
    181  * factors are defined in the timex.h header file.
    182  *
    183  * time_phase and time_freq are the phase increment and the frequency
    184  * increment, respectively, of the kernel time variable.
    185  *
    186  * time_freq is set via ntp_adjtime() from a value stored in a file when
    187  * the synchronization daemon is first started. Its value is retrieved
    188  * via ntp_adjtime() and written to the file about once per hour by the
    189  * daemon.
    190  *
    191  * time_adj is the adjustment added to the value of tick at each timer
    192  * interrupt and is recomputed from time_phase and time_freq at each
    193  * seconds rollover.
    194  *
    195  * time_reftime is the second's portion of the system time at the last
    196  * call to ntp_adjtime(). It is used to adjust the time_freq variable
    197  * and to increase the time_maxerror as the time since last update
    198  * increases.
    199  */
    200 long time_phase = 0;		/* phase offset (scaled us) */
    201 long time_freq = 0;		/* frequency offset (scaled ppm) */
    202 long time_adj = 0;		/* tick adjust (scaled 1 / hz) */
    203 long time_reftime = 0;		/* time at last adjustment (s) */
    204 
    205 #ifdef PPS_SYNC
    206 /*
    207  * The following variables are used only if the kernel PPS discipline
    208  * code is configured (PPS_SYNC). The scale factors are defined in the
    209  * timex.h header file.
    210  *
    211  * pps_time contains the time at each calibration interval, as read by
    212  * microtime(). pps_count counts the seconds of the calibration
    213  * interval, the duration of which is nominally pps_shift in powers of
    214  * two.
    215  *
    216  * pps_offset is the time offset produced by the time median filter
    217  * pps_tf[], while pps_jitter is the dispersion (jitter) measured by
    218  * this filter.
    219  *
    220  * pps_freq is the frequency offset produced by the frequency median
    221  * filter pps_ff[], while pps_stabil is the dispersion (wander) measured
    222  * by this filter.
    223  *
    224  * pps_usec is latched from a high resolution counter or external clock
    225  * at pps_time. Here we want the hardware counter contents only, not the
    226  * contents plus the time_tv.usec as usual.
    227  *
    228  * pps_valid counts the number of seconds since the last PPS update. It
    229  * is used as a watchdog timer to disable the PPS discipline should the
    230  * PPS signal be lost.
    231  *
    232  * pps_glitch counts the number of seconds since the beginning of an
    233  * offset burst more than tick/2 from current nominal offset. It is used
    234  * mainly to suppress error bursts due to priority conflicts between the
    235  * PPS interrupt and timer interrupt.
    236  *
    237  * pps_intcnt counts the calibration intervals for use in the interval-
    238  * adaptation algorithm. It's just too complicated for words.
    239  *
    240  * pps_kc_hardpps_source contains an arbitrary value that uniquely
    241  * identifies the currently bound source of the PPS signal, or NULL
    242  * if no source is bound.
    243  *
    244  * pps_kc_hardpps_mode indicates which transitions, if any, of the PPS
    245  * signal should be reported.
    246  */
    247 struct timeval pps_time;	/* kernel time at last interval */
    248 long pps_tf[] = {0, 0, 0};	/* pps time offset median filter (us) */
    249 long pps_offset = 0;		/* pps time offset (us) */
    250 long pps_jitter = MAXTIME;	/* time dispersion (jitter) (us) */
    251 long pps_ff[] = {0, 0, 0};	/* pps frequency offset median filter */
    252 long pps_freq = 0;		/* frequency offset (scaled ppm) */
    253 long pps_stabil = MAXFREQ;	/* frequency dispersion (scaled ppm) */
    254 long pps_usec = 0;		/* microsec counter at last interval */
    255 long pps_valid = PPS_VALID;	/* pps signal watchdog counter */
    256 int pps_glitch = 0;		/* pps signal glitch counter */
    257 int pps_count = 0;		/* calibration interval counter (s) */
    258 int pps_shift = PPS_SHIFT;	/* interval duration (s) (shift) */
    259 int pps_intcnt = 0;		/* intervals at current duration */
    260 void *pps_kc_hardpps_source = NULL; /* current PPS supplier's identifier */
    261 int pps_kc_hardpps_mode = 0;	/* interesting edges of PPS signal */
    262 
    263 /*
    264  * PPS signal quality monitors
    265  *
    266  * pps_jitcnt counts the seconds that have been discarded because the
    267  * jitter measured by the time median filter exceeds the limit MAXTIME
    268  * (100 us).
    269  *
    270  * pps_calcnt counts the frequency calibration intervals, which are
    271  * variable from 4 s to 256 s.
    272  *
    273  * pps_errcnt counts the calibration intervals which have been discarded
    274  * because the wander exceeds the limit MAXFREQ (100 ppm) or where the
    275  * calibration interval jitter exceeds two ticks.
    276  *
    277  * pps_stbcnt counts the calibration intervals that have been discarded
    278  * because the frequency wander exceeds the limit MAXFREQ / 4 (25 us).
    279  */
    280 long pps_jitcnt = 0;		/* jitter limit exceeded */
    281 long pps_calcnt = 0;		/* calibration intervals */
    282 long pps_errcnt = 0;		/* calibration errors */
    283 long pps_stbcnt = 0;		/* stability limit exceeded */
    284 #endif /* PPS_SYNC */
    285 
    286 #ifdef EXT_CLOCK
    287 /*
    288  * External clock definitions
    289  *
    290  * The following definitions and declarations are used only if an
    291  * external clock is configured on the system.
    292  */
    293 #define CLOCK_INTERVAL 30	/* CPU clock update interval (s) */
    294 
    295 /*
    296  * The clock_count variable is set to CLOCK_INTERVAL at each PPS
    297  * interrupt and decremented once each second.
    298  */
    299 int clock_count = 0;		/* CPU clock counter */
    300 
    301 #ifdef HIGHBALL
    302 /*
    303  * The clock_offset and clock_cpu variables are used by the HIGHBALL
    304  * interface. The clock_offset variable defines the offset between
    305  * system time and the HIGBALL counters. The clock_cpu variable contains
    306  * the offset between the system clock and the HIGHBALL clock for use in
    307  * disciplining the kernel time variable.
    308  */
    309 extern struct timeval clock_offset; /* Highball clock offset */
    310 long clock_cpu = 0;		/* CPU clock adjust */
    311 #endif /* HIGHBALL */
    312 #endif /* EXT_CLOCK */
    313 #endif /* NTP */
    314 
    315 /*
    316  * Bump a timeval by a small number of usec's.
    317  */
    318 #define BUMPTIME(t, usec) { \
    319 	volatile struct timeval *tp = (t); \
    320 	long us; \
    321  \
    322 	tp->tv_usec = us = tp->tv_usec + (usec); \
    323 	if (us >= 1000000) { \
    324 		tp->tv_usec = us - 1000000; \
    325 		tp->tv_sec++; \
    326 	} \
    327 }
    328 #endif /* !__HAVE_TIMECOUNTER */
    329 
    330 int	stathz;
    331 int	profhz;
    332 int	profsrc;
    333 int	schedhz;
    334 int	profprocs;
    335 int	hardclock_ticks;
    336 static int statscheddiv; /* stat => sched divider (used if schedhz == 0) */
    337 static int psdiv;			/* prof => stat divider */
    338 int	psratio;			/* ratio: prof / stat */
    339 #ifndef __HAVE_TIMECOUNTER
    340 int	tickfix, tickfixinterval;	/* used if tick not really integral */
    341 #ifndef NTP
    342 static int tickfixcnt;			/* accumulated fractional error */
    343 #else
    344 int	fixtick;			/* used by NTP for same */
    345 int	shifthz;
    346 #endif
    347 
    348 /*
    349  * We might want ldd to load the both words from time at once.
    350  * To succeed we need to be quadword aligned.
    351  * The sparc already does that, and that it has worked so far is a fluke.
    352  */
    353 volatile struct	timeval time  __attribute__((__aligned__(__alignof__(quad_t))));
    354 volatile struct	timeval mono_time;
    355 #endif /* !__HAVE_TIMECOUNTER */
    356 
    357 #ifdef __HAVE_GENERIC_SOFT_INTERRUPTS
    358 void	*softclock_si;
    359 #endif
    360 
    361 #ifdef __HAVE_TIMECOUNTER
    362 static u_int get_intr_timecount(struct timecounter *);
    363 
    364 static struct timecounter intr_timecounter = {
    365 	get_intr_timecount,	/* get_timecount */
    366 	0,			/* no poll_pps */
    367 	~0u,			/* counter_mask */
    368 	0,		        /* frequency */
    369 	"clockinterrupt",	/* name */
    370 	0,			/* quality - minimum implementation level for a clock */
    371 	NULL,			/* prev */
    372 	NULL,			/* next */
    373 };
    374 
    375 static u_int
    376 get_intr_timecount(struct timecounter *tc)
    377 {
    378 
    379 	return (u_int)hardclock_ticks;
    380 }
    381 #endif
    382 
    383 /*
    384  * Initialize clock frequencies and start both clocks running.
    385  */
    386 void
    387 initclocks(void)
    388 {
    389 	int i;
    390 
    391 #ifdef __HAVE_GENERIC_SOFT_INTERRUPTS
    392 	softclock_si = softintr_establish(IPL_SOFTCLOCK, softclock, NULL);
    393 	if (softclock_si == NULL)
    394 		panic("initclocks: unable to register softclock intr");
    395 #endif
    396 
    397 	/*
    398 	 * Set divisors to 1 (normal case) and let the machine-specific
    399 	 * code do its bit.
    400 	 */
    401 	psdiv = 1;
    402 #ifdef __HAVE_TIMECOUNTER
    403 	/*
    404 	 * provide minimum default time counter
    405 	 * will only run at interrupt resolution
    406 	 */
    407 	intr_timecounter.tc_frequency = hz;
    408 	tc_init(&intr_timecounter);
    409 #endif
    410 	cpu_initclocks();
    411 
    412 	/*
    413 	 * Compute profhz/stathz/rrticks, and fix profhz if needed.
    414 	 */
    415 	i = stathz ? stathz : hz;
    416 	if (profhz == 0)
    417 		profhz = i;
    418 	psratio = profhz / i;
    419 	rrticks = hz / 10;
    420 	if (schedhz == 0) {
    421 		/* 16Hz is best */
    422 		statscheddiv = i / 16;
    423 		if (statscheddiv <= 0)
    424 			panic("statscheddiv");
    425 	}
    426 
    427 #ifndef __HAVE_TIMECOUNTER
    428 #ifdef NTP
    429 	switch (hz) {
    430 	case 1:
    431 		shifthz = SHIFT_SCALE - 0;
    432 		break;
    433 	case 2:
    434 		shifthz = SHIFT_SCALE - 1;
    435 		break;
    436 	case 4:
    437 		shifthz = SHIFT_SCALE - 2;
    438 		break;
    439 	case 8:
    440 		shifthz = SHIFT_SCALE - 3;
    441 		break;
    442 	case 16:
    443 		shifthz = SHIFT_SCALE - 4;
    444 		break;
    445 	case 32:
    446 		shifthz = SHIFT_SCALE - 5;
    447 		break;
    448 	case 50:
    449 	case 60:
    450 	case 64:
    451 		shifthz = SHIFT_SCALE - 6;
    452 		break;
    453 	case 96:
    454 	case 100:
    455 	case 128:
    456 		shifthz = SHIFT_SCALE - 7;
    457 		break;
    458 	case 256:
    459 		shifthz = SHIFT_SCALE - 8;
    460 		break;
    461 	case 512:
    462 		shifthz = SHIFT_SCALE - 9;
    463 		break;
    464 	case 1000:
    465 	case 1024:
    466 		shifthz = SHIFT_SCALE - 10;
    467 		break;
    468 	case 1200:
    469 	case 2048:
    470 		shifthz = SHIFT_SCALE - 11;
    471 		break;
    472 	case 4096:
    473 		shifthz = SHIFT_SCALE - 12;
    474 		break;
    475 	case 8192:
    476 		shifthz = SHIFT_SCALE - 13;
    477 		break;
    478 	case 16384:
    479 		shifthz = SHIFT_SCALE - 14;
    480 		break;
    481 	case 32768:
    482 		shifthz = SHIFT_SCALE - 15;
    483 		break;
    484 	case 65536:
    485 		shifthz = SHIFT_SCALE - 16;
    486 		break;
    487 	default:
    488 		panic("weird hz");
    489 	}
    490 	if (fixtick == 0) {
    491 		/*
    492 		 * Give MD code a chance to set this to a better
    493 		 * value; but, if it doesn't, we should.
    494 		 */
    495 		fixtick = (1000000 - (hz*tick));
    496 	}
    497 #endif /* NTP */
    498 #endif /* !__HAVE_TIMECOUNTER */
    499 }
    500 
    501 /*
    502  * The real-time timer, interrupting hz times per second.
    503  */
    504 void
    505 hardclock(struct clockframe *frame)
    506 {
    507 	struct lwp *l;
    508 	struct proc *p;
    509 	struct cpu_info *ci = curcpu();
    510 	struct ptimer *pt;
    511 #ifndef __HAVE_TIMECOUNTER
    512 	int delta;
    513 	extern int tickdelta;
    514 	extern long timedelta;
    515 #ifdef NTP
    516 	int time_update;
    517 	int ltemp;
    518 #endif /* NTP */
    519 #endif /* __HAVE_TIMECOUNTER */
    520 
    521 	l = curlwp;
    522 	if (l) {
    523 		p = l->l_proc;
    524 		/*
    525 		 * Run current process's virtual and profile time, as needed.
    526 		 */
    527 		if (CLKF_USERMODE(frame) && p->p_timers &&
    528 		    (pt = LIST_FIRST(&p->p_timers->pts_virtual)) != NULL)
    529 			if (itimerdecr(pt, tick) == 0)
    530 				itimerfire(pt);
    531 		if (p->p_timers &&
    532 		    (pt = LIST_FIRST(&p->p_timers->pts_prof)) != NULL)
    533 			if (itimerdecr(pt, tick) == 0)
    534 				itimerfire(pt);
    535 	}
    536 
    537 	/*
    538 	 * If no separate statistics clock is available, run it from here.
    539 	 */
    540 	if (stathz == 0)
    541 		statclock(frame);
    542 	if ((--ci->ci_schedstate.spc_rrticks) <= 0)
    543 		roundrobin(ci);
    544 
    545 #if defined(MULTIPROCESSOR)
    546 	/*
    547 	 * If we are not the primary CPU, we're not allowed to do
    548 	 * any more work.
    549 	 */
    550 	if (CPU_IS_PRIMARY(ci) == 0)
    551 		return;
    552 #endif
    553 
    554 	hardclock_ticks++;
    555 
    556 #ifdef __HAVE_TIMECOUNTER
    557 	tc_ticktock();
    558 #else /* __HAVE_TIMECOUNTER */
    559 	/*
    560 	 * Increment the time-of-day.  The increment is normally just
    561 	 * ``tick''.  If the machine is one which has a clock frequency
    562 	 * such that ``hz'' would not divide the second evenly into
    563 	 * milliseconds, a periodic adjustment must be applied.  Finally,
    564 	 * if we are still adjusting the time (see adjtime()),
    565 	 * ``tickdelta'' may also be added in.
    566 	 */
    567 	delta = tick;
    568 
    569 #ifndef NTP
    570 	if (tickfix) {
    571 		tickfixcnt += tickfix;
    572 		if (tickfixcnt >= tickfixinterval) {
    573 			delta++;
    574 			tickfixcnt -= tickfixinterval;
    575 		}
    576 	}
    577 #endif /* !NTP */
    578 	/* Imprecise 4bsd adjtime() handling */
    579 	if (timedelta != 0) {
    580 		delta += tickdelta;
    581 		timedelta -= tickdelta;
    582 	}
    583 
    584 #ifdef notyet
    585 	microset();
    586 #endif
    587 
    588 #ifndef NTP
    589 	BUMPTIME(&time, delta);		/* XXX Now done using NTP code below */
    590 #endif
    591 	BUMPTIME(&mono_time, delta);
    592 
    593 #ifdef NTP
    594 	time_update = delta;
    595 
    596 	/*
    597 	 * Compute the phase adjustment. If the low-order bits
    598 	 * (time_phase) of the update overflow, bump the high-order bits
    599 	 * (time_update).
    600 	 */
    601 	time_phase += time_adj;
    602 	if (time_phase <= -FINEUSEC) {
    603 		ltemp = -time_phase >> SHIFT_SCALE;
    604 		time_phase += ltemp << SHIFT_SCALE;
    605 		time_update -= ltemp;
    606 	} else if (time_phase >= FINEUSEC) {
    607 		ltemp = time_phase >> SHIFT_SCALE;
    608 		time_phase -= ltemp << SHIFT_SCALE;
    609 		time_update += ltemp;
    610 	}
    611 
    612 #ifdef HIGHBALL
    613 	/*
    614 	 * If the HIGHBALL board is installed, we need to adjust the
    615 	 * external clock offset in order to close the hardware feedback
    616 	 * loop. This will adjust the external clock phase and frequency
    617 	 * in small amounts. The additional phase noise and frequency
    618 	 * wander this causes should be minimal. We also need to
    619 	 * discipline the kernel time variable, since the PLL is used to
    620 	 * discipline the external clock. If the Highball board is not
    621 	 * present, we discipline kernel time with the PLL as usual. We
    622 	 * assume that the external clock phase adjustment (time_update)
    623 	 * and kernel phase adjustment (clock_cpu) are less than the
    624 	 * value of tick.
    625 	 */
    626 	clock_offset.tv_usec += time_update;
    627 	if (clock_offset.tv_usec >= 1000000) {
    628 		clock_offset.tv_sec++;
    629 		clock_offset.tv_usec -= 1000000;
    630 	}
    631 	if (clock_offset.tv_usec < 0) {
    632 		clock_offset.tv_sec--;
    633 		clock_offset.tv_usec += 1000000;
    634 	}
    635 	time.tv_usec += clock_cpu;
    636 	clock_cpu = 0;
    637 #else
    638 	time.tv_usec += time_update;
    639 #endif /* HIGHBALL */
    640 
    641 	/*
    642 	 * On rollover of the second the phase adjustment to be used for
    643 	 * the next second is calculated. Also, the maximum error is
    644 	 * increased by the tolerance. If the PPS frequency discipline
    645 	 * code is present, the phase is increased to compensate for the
    646 	 * CPU clock oscillator frequency error.
    647 	 *
    648  	 * On a 32-bit machine and given parameters in the timex.h
    649 	 * header file, the maximum phase adjustment is +-512 ms and
    650 	 * maximum frequency offset is a tad less than) +-512 ppm. On a
    651 	 * 64-bit machine, you shouldn't need to ask.
    652 	 */
    653 	if (time.tv_usec >= 1000000) {
    654 		time.tv_usec -= 1000000;
    655 		time.tv_sec++;
    656 		time_maxerror += time_tolerance >> SHIFT_USEC;
    657 
    658 		/*
    659 		 * Leap second processing. If in leap-insert state at
    660 		 * the end of the day, the system clock is set back one
    661 		 * second; if in leap-delete state, the system clock is
    662 		 * set ahead one second. The microtime() routine or
    663 		 * external clock driver will insure that reported time
    664 		 * is always monotonic. The ugly divides should be
    665 		 * replaced.
    666 		 */
    667 		switch (time_state) {
    668 		case TIME_OK:
    669 			if (time_status & STA_INS)
    670 				time_state = TIME_INS;
    671 			else if (time_status & STA_DEL)
    672 				time_state = TIME_DEL;
    673 			break;
    674 
    675 		case TIME_INS:
    676 			if (time.tv_sec % 86400 == 0) {
    677 				time.tv_sec--;
    678 				time_state = TIME_OOP;
    679 			}
    680 			break;
    681 
    682 		case TIME_DEL:
    683 			if ((time.tv_sec + 1) % 86400 == 0) {
    684 				time.tv_sec++;
    685 				time_state = TIME_WAIT;
    686 			}
    687 			break;
    688 
    689 		case TIME_OOP:
    690 			time_state = TIME_WAIT;
    691 			break;
    692 
    693 		case TIME_WAIT:
    694 			if (!(time_status & (STA_INS | STA_DEL)))
    695 				time_state = TIME_OK;
    696 			break;
    697 		}
    698 
    699 		/*
    700 		 * Compute the phase adjustment for the next second. In
    701 		 * PLL mode, the offset is reduced by a fixed factor
    702 		 * times the time constant. In FLL mode the offset is
    703 		 * used directly. In either mode, the maximum phase
    704 		 * adjustment for each second is clamped so as to spread
    705 		 * the adjustment over not more than the number of
    706 		 * seconds between updates.
    707 		 */
    708 		if (time_offset < 0) {
    709 			ltemp = -time_offset;
    710 			if (!(time_status & STA_FLL))
    711 				ltemp >>= SHIFT_KG + time_constant;
    712 			if (ltemp > (MAXPHASE / MINSEC) << SHIFT_UPDATE)
    713 				ltemp = (MAXPHASE / MINSEC) <<
    714 				    SHIFT_UPDATE;
    715 			time_offset += ltemp;
    716 			time_adj = -ltemp << (shifthz - SHIFT_UPDATE);
    717 		} else if (time_offset > 0) {
    718 			ltemp = time_offset;
    719 			if (!(time_status & STA_FLL))
    720 				ltemp >>= SHIFT_KG + time_constant;
    721 			if (ltemp > (MAXPHASE / MINSEC) << SHIFT_UPDATE)
    722 				ltemp = (MAXPHASE / MINSEC) <<
    723 				    SHIFT_UPDATE;
    724 			time_offset -= ltemp;
    725 			time_adj = ltemp << (shifthz - SHIFT_UPDATE);
    726 		} else
    727 			time_adj = 0;
    728 
    729 		/*
    730 		 * Compute the frequency estimate and additional phase
    731 		 * adjustment due to frequency error for the next
    732 		 * second. When the PPS signal is engaged, gnaw on the
    733 		 * watchdog counter and update the frequency computed by
    734 		 * the pll and the PPS signal.
    735 		 */
    736 #ifdef PPS_SYNC
    737 		pps_valid++;
    738 		if (pps_valid == PPS_VALID) {
    739 			pps_jitter = MAXTIME;
    740 			pps_stabil = MAXFREQ;
    741 			time_status &= ~(STA_PPSSIGNAL | STA_PPSJITTER |
    742 			    STA_PPSWANDER | STA_PPSERROR);
    743 		}
    744 		ltemp = time_freq + pps_freq;
    745 #else
    746 		ltemp = time_freq;
    747 #endif /* PPS_SYNC */
    748 
    749 		if (ltemp < 0)
    750 			time_adj -= -ltemp >> (SHIFT_USEC - shifthz);
    751 		else
    752 			time_adj += ltemp >> (SHIFT_USEC - shifthz);
    753 		time_adj += (long)fixtick << shifthz;
    754 
    755 		/*
    756 		 * When the CPU clock oscillator frequency is not a
    757 		 * power of 2 in Hz, shifthz is only an approximate
    758 		 * scale factor.
    759 		 *
    760 		 * To determine the adjustment, you can do the following:
    761 		 *   bc -q
    762 		 *   scale=24
    763 		 *   obase=2
    764 		 *   idealhz/realhz
    765 		 * where `idealhz' is the next higher power of 2, and `realhz'
    766 		 * is the actual value.  You may need to factor this result
    767 		 * into a sequence of 2 multipliers to get better precision.
    768 		 *
    769 		 * Likewise, the error can be calculated with (e.g. for 100Hz):
    770 		 *   bc -q
    771 		 *   scale=24
    772 		 *   ((1+2^-2+2^-5)*(1-2^-10)*realhz-idealhz)/idealhz
    773 		 * (and then multiply by 1000000 to get ppm).
    774 		 */
    775 		switch (hz) {
    776 		case 60:
    777 			/* A factor of 1.000100010001 gives about 15ppm
    778 			   error. */
    779 			if (time_adj < 0) {
    780 				time_adj -= (-time_adj >> 4);
    781 				time_adj -= (-time_adj >> 8);
    782 			} else {
    783 				time_adj += (time_adj >> 4);
    784 				time_adj += (time_adj >> 8);
    785 			}
    786 			break;
    787 
    788 		case 96:
    789 			/* A factor of 1.0101010101 gives about 244ppm error. */
    790 			if (time_adj < 0) {
    791 				time_adj -= (-time_adj >> 2);
    792 				time_adj -= (-time_adj >> 4) + (-time_adj >> 8);
    793 			} else {
    794 				time_adj += (time_adj >> 2);
    795 				time_adj += (time_adj >> 4) + (time_adj >> 8);
    796 			}
    797 			break;
    798 
    799 		case 50:
    800 		case 100:
    801 			/* A factor of 1.010001111010111 gives about 1ppm
    802 			   error. */
    803 			if (time_adj < 0) {
    804 				time_adj -= (-time_adj >> 2) + (-time_adj >> 5);
    805 				time_adj += (-time_adj >> 10);
    806 			} else {
    807 				time_adj += (time_adj >> 2) + (time_adj >> 5);
    808 				time_adj -= (time_adj >> 10);
    809 			}
    810 			break;
    811 
    812 		case 1000:
    813 			/* A factor of 1.000001100010100001 gives about 50ppm
    814 			   error. */
    815 			if (time_adj < 0) {
    816 				time_adj -= (-time_adj >> 6) + (-time_adj >> 11);
    817 				time_adj -= (-time_adj >> 7);
    818 			} else {
    819 				time_adj += (time_adj >> 6) + (time_adj >> 11);
    820 				time_adj += (time_adj >> 7);
    821 			}
    822 			break;
    823 
    824 		case 1200:
    825 			/* A factor of 1.1011010011100001 gives about 64ppm
    826 			   error. */
    827 			if (time_adj < 0) {
    828 				time_adj -= (-time_adj >> 1) + (-time_adj >> 6);
    829 				time_adj -= (-time_adj >> 3) + (-time_adj >> 10);
    830 			} else {
    831 				time_adj += (time_adj >> 1) + (time_adj >> 6);
    832 				time_adj += (time_adj >> 3) + (time_adj >> 10);
    833 			}
    834 			break;
    835 		}
    836 
    837 #ifdef EXT_CLOCK
    838 		/*
    839 		 * If an external clock is present, it is necessary to
    840 		 * discipline the kernel time variable anyway, since not
    841 		 * all system components use the microtime() interface.
    842 		 * Here, the time offset between the external clock and
    843 		 * kernel time variable is computed every so often.
    844 		 */
    845 		clock_count++;
    846 		if (clock_count > CLOCK_INTERVAL) {
    847 			clock_count = 0;
    848 			microtime(&clock_ext);
    849 			delta.tv_sec = clock_ext.tv_sec - time.tv_sec;
    850 			delta.tv_usec = clock_ext.tv_usec -
    851 			    time.tv_usec;
    852 			if (delta.tv_usec < 0)
    853 				delta.tv_sec--;
    854 			if (delta.tv_usec >= 500000) {
    855 				delta.tv_usec -= 1000000;
    856 				delta.tv_sec++;
    857 			}
    858 			if (delta.tv_usec < -500000) {
    859 				delta.tv_usec += 1000000;
    860 				delta.tv_sec--;
    861 			}
    862 			if (delta.tv_sec > 0 || (delta.tv_sec == 0 &&
    863 			    delta.tv_usec > MAXPHASE) ||
    864 			    delta.tv_sec < -1 || (delta.tv_sec == -1 &&
    865 			    delta.tv_usec < -MAXPHASE)) {
    866 				time = clock_ext;
    867 				delta.tv_sec = 0;
    868 				delta.tv_usec = 0;
    869 			}
    870 #ifdef HIGHBALL
    871 			clock_cpu = delta.tv_usec;
    872 #else /* HIGHBALL */
    873 			hardupdate(delta.tv_usec);
    874 #endif /* HIGHBALL */
    875 		}
    876 #endif /* EXT_CLOCK */
    877 	}
    878 
    879 #endif /* NTP */
    880 #endif /* !__HAVE_TIMECOUNTER */
    881 
    882 	/*
    883 	 * Update real-time timeout queue.
    884 	 * Process callouts at a very low CPU priority, so we don't keep the
    885 	 * relatively high clock interrupt priority any longer than necessary.
    886 	 */
    887 	if (callout_hardclock()) {
    888 		if (CLKF_BASEPRI(frame)) {
    889 			/*
    890 			 * Save the overhead of a software interrupt;
    891 			 * it will happen as soon as we return, so do
    892 			 * it now.
    893 			 */
    894 			spllowersoftclock();
    895 			KERNEL_LOCK(LK_CANRECURSE|LK_EXCLUSIVE);
    896 			softclock(NULL);
    897 			KERNEL_UNLOCK();
    898 		} else {
    899 #ifdef __HAVE_GENERIC_SOFT_INTERRUPTS
    900 			softintr_schedule(softclock_si);
    901 #else
    902 			setsoftclock();
    903 #endif
    904 		}
    905 	}
    906 }
    907 
    908 #ifdef __HAVE_TIMECOUNTER
    909 /*
    910  * Compute number of hz until specified time.  Used to compute second
    911  * argument to callout_reset() from an absolute time.
    912  */
    913 int
    914 hzto(struct timeval *tvp)
    915 {
    916 	struct timeval now, tv;
    917 
    918 	tv = *tvp;	/* Don't modify original tvp. */
    919 	getmicrotime(&now);
    920 	timersub(&tv, &now, &tv);
    921 	return tvtohz(&tv);
    922 }
    923 #endif /* __HAVE_TIMECOUNTER */
    924 
    925 /*
    926  * Compute number of ticks in the specified amount of time.
    927  */
    928 int
    929 tvtohz(struct timeval *tv)
    930 {
    931 	unsigned long ticks;
    932 	long sec, usec;
    933 
    934 	/*
    935 	 * If the number of usecs in the whole seconds part of the time
    936 	 * difference fits in a long, then the total number of usecs will
    937 	 * fit in an unsigned long.  Compute the total and convert it to
    938 	 * ticks, rounding up and adding 1 to allow for the current tick
    939 	 * to expire.  Rounding also depends on unsigned long arithmetic
    940 	 * to avoid overflow.
    941 	 *
    942 	 * Otherwise, if the number of ticks in the whole seconds part of
    943 	 * the time difference fits in a long, then convert the parts to
    944 	 * ticks separately and add, using similar rounding methods and
    945 	 * overflow avoidance.  This method would work in the previous
    946 	 * case, but it is slightly slower and assumes that hz is integral.
    947 	 *
    948 	 * Otherwise, round the time difference down to the maximum
    949 	 * representable value.
    950 	 *
    951 	 * If ints are 32-bit, then the maximum value for any timeout in
    952 	 * 10ms ticks is 248 days.
    953 	 */
    954 	sec = tv->tv_sec;
    955 	usec = tv->tv_usec;
    956 
    957 	if (usec < 0) {
    958 		sec--;
    959 		usec += 1000000;
    960 	}
    961 
    962 	if (sec < 0 || (sec == 0 && usec <= 0)) {
    963 		/*
    964 		 * Would expire now or in the past.  Return 0 ticks.
    965 		 * This is different from the legacy hzto() interface,
    966 		 * and callers need to check for it.
    967 		 */
    968 		ticks = 0;
    969 	} else if (sec <= (LONG_MAX / 1000000))
    970 		ticks = (((sec * 1000000) + (unsigned long)usec + (tick - 1))
    971 		    / tick) + 1;
    972 	else if (sec <= (LONG_MAX / hz))
    973 		ticks = (sec * hz) +
    974 		    (((unsigned long)usec + (tick - 1)) / tick) + 1;
    975 	else
    976 		ticks = LONG_MAX;
    977 
    978 	if (ticks > INT_MAX)
    979 		ticks = INT_MAX;
    980 
    981 	return ((int)ticks);
    982 }
    983 
    984 #ifndef __HAVE_TIMECOUNTER
    985 /*
    986  * Compute number of hz until specified time.  Used to compute second
    987  * argument to callout_reset() from an absolute time.
    988  */
    989 int
    990 hzto(struct timeval *tv)
    991 {
    992 	unsigned long ticks;
    993 	long sec, usec;
    994 	int s;
    995 
    996 	/*
    997 	 * If the number of usecs in the whole seconds part of the time
    998 	 * difference fits in a long, then the total number of usecs will
    999 	 * fit in an unsigned long.  Compute the total and convert it to
   1000 	 * ticks, rounding up and adding 1 to allow for the current tick
   1001 	 * to expire.  Rounding also depends on unsigned long arithmetic
   1002 	 * to avoid overflow.
   1003 	 *
   1004 	 * Otherwise, if the number of ticks in the whole seconds part of
   1005 	 * the time difference fits in a long, then convert the parts to
   1006 	 * ticks separately and add, using similar rounding methods and
   1007 	 * overflow avoidance.  This method would work in the previous
   1008 	 * case, but it is slightly slower and assume that hz is integral.
   1009 	 *
   1010 	 * Otherwise, round the time difference down to the maximum
   1011 	 * representable value.
   1012 	 *
   1013 	 * If ints are 32-bit, then the maximum value for any timeout in
   1014 	 * 10ms ticks is 248 days.
   1015 	 */
   1016 	s = splclock();
   1017 	sec = tv->tv_sec - time.tv_sec;
   1018 	usec = tv->tv_usec - time.tv_usec;
   1019 	splx(s);
   1020 
   1021 	if (usec < 0) {
   1022 		sec--;
   1023 		usec += 1000000;
   1024 	}
   1025 
   1026 	if (sec < 0 || (sec == 0 && usec <= 0)) {
   1027 		/*
   1028 		 * Would expire now or in the past.  Return 0 ticks.
   1029 		 * This is different from the legacy hzto() interface,
   1030 		 * and callers need to check for it.
   1031 		 */
   1032 		ticks = 0;
   1033 	} else if (sec <= (LONG_MAX / 1000000))
   1034 		ticks = (((sec * 1000000) + (unsigned long)usec + (tick - 1))
   1035 		    / tick) + 1;
   1036 	else if (sec <= (LONG_MAX / hz))
   1037 		ticks = (sec * hz) +
   1038 		    (((unsigned long)usec + (tick - 1)) / tick) + 1;
   1039 	else
   1040 		ticks = LONG_MAX;
   1041 
   1042 	if (ticks > INT_MAX)
   1043 		ticks = INT_MAX;
   1044 
   1045 	return ((int)ticks);
   1046 }
   1047 #endif /* !__HAVE_TIMECOUNTER */
   1048 
   1049 /*
   1050  * Compute number of ticks in the specified amount of time.
   1051  */
   1052 int
   1053 tstohz(struct timespec *ts)
   1054 {
   1055 	struct timeval tv;
   1056 
   1057 	/*
   1058 	 * usec has great enough resolution for hz, so convert to a
   1059 	 * timeval and use tvtohz() above.
   1060 	 */
   1061 	TIMESPEC_TO_TIMEVAL(&tv, ts);
   1062 	return tvtohz(&tv);
   1063 }
   1064 
   1065 /*
   1066  * Start profiling on a process.
   1067  *
   1068  * Kernel profiling passes proc0 which never exits and hence
   1069  * keeps the profile clock running constantly.
   1070  */
   1071 void
   1072 startprofclock(struct proc *p)
   1073 {
   1074 
   1075 	if ((p->p_flag & P_PROFIL) == 0) {
   1076 		p->p_flag |= P_PROFIL;
   1077 		/*
   1078 		 * This is only necessary if using the clock as the
   1079 		 * profiling source.
   1080 		 */
   1081 		if (++profprocs == 1 && stathz != 0)
   1082 			psdiv = psratio;
   1083 	}
   1084 }
   1085 
   1086 /*
   1087  * Stop profiling on a process.
   1088  */
   1089 void
   1090 stopprofclock(struct proc *p)
   1091 {
   1092 
   1093 	if (p->p_flag & P_PROFIL) {
   1094 		p->p_flag &= ~P_PROFIL;
   1095 		/*
   1096 		 * This is only necessary if using the clock as the
   1097 		 * profiling source.
   1098 		 */
   1099 		if (--profprocs == 0 && stathz != 0)
   1100 			psdiv = 1;
   1101 	}
   1102 }
   1103 
   1104 #if defined(PERFCTRS)
   1105 /*
   1106  * Independent profiling "tick" in case we're using a separate
   1107  * clock or profiling event source.  Currently, that's just
   1108  * performance counters--hence the wrapper.
   1109  */
   1110 void
   1111 proftick(struct clockframe *frame)
   1112 {
   1113 #ifdef GPROF
   1114         struct gmonparam *g;
   1115         intptr_t i;
   1116 #endif
   1117 	struct proc *p;
   1118 
   1119 	p = curproc;
   1120 	if (CLKF_USERMODE(frame)) {
   1121 		if (p->p_flag & P_PROFIL)
   1122 			addupc_intr(p, CLKF_PC(frame));
   1123 	} else {
   1124 #ifdef GPROF
   1125 		g = &_gmonparam;
   1126 		if (g->state == GMON_PROF_ON) {
   1127 			i = CLKF_PC(frame) - g->lowpc;
   1128 			if (i < g->textsize) {
   1129 				i /= HISTFRACTION * sizeof(*g->kcount);
   1130 				g->kcount[i]++;
   1131 			}
   1132 		}
   1133 #endif
   1134 #ifdef PROC_PC
   1135                 if (p && (p->p_flag & P_PROFIL))
   1136                         addupc_intr(p, PROC_PC(p));
   1137 #endif
   1138 	}
   1139 }
   1140 #endif
   1141 
   1142 /*
   1143  * Statistics clock.  Grab profile sample, and if divider reaches 0,
   1144  * do process and kernel statistics.
   1145  */
   1146 void
   1147 statclock(struct clockframe *frame)
   1148 {
   1149 #ifdef GPROF
   1150 	struct gmonparam *g;
   1151 	intptr_t i;
   1152 #endif
   1153 	struct cpu_info *ci = curcpu();
   1154 	struct schedstate_percpu *spc = &ci->ci_schedstate;
   1155 	struct proc *p;
   1156 	struct lwp *l;
   1157 
   1158 	/*
   1159 	 * Notice changes in divisor frequency, and adjust clock
   1160 	 * frequency accordingly.
   1161 	 */
   1162 	if (spc->spc_psdiv != psdiv) {
   1163 		spc->spc_psdiv = psdiv;
   1164 		spc->spc_pscnt = psdiv;
   1165 		if (psdiv == 1) {
   1166 			setstatclockrate(stathz);
   1167 		} else {
   1168 			setstatclockrate(profhz);
   1169 		}
   1170 	}
   1171 	l = curlwp;
   1172 	p = (l ? l->l_proc : NULL);
   1173 	if (CLKF_USERMODE(frame)) {
   1174 		KASSERT(p != NULL);
   1175 
   1176 		if ((p->p_flag & P_PROFIL) && profsrc == PROFSRC_CLOCK)
   1177 			addupc_intr(p, CLKF_PC(frame));
   1178 		if (--spc->spc_pscnt > 0)
   1179 			return;
   1180 		/*
   1181 		 * Came from user mode; CPU was in user state.
   1182 		 * If this process is being profiled record the tick.
   1183 		 */
   1184 		p->p_uticks++;
   1185 		if (p->p_nice > NZERO)
   1186 			spc->spc_cp_time[CP_NICE]++;
   1187 		else
   1188 			spc->spc_cp_time[CP_USER]++;
   1189 	} else {
   1190 #ifdef GPROF
   1191 		/*
   1192 		 * Kernel statistics are just like addupc_intr, only easier.
   1193 		 */
   1194 		g = &_gmonparam;
   1195 		if (profsrc == PROFSRC_CLOCK && g->state == GMON_PROF_ON) {
   1196 			i = CLKF_PC(frame) - g->lowpc;
   1197 			if (i < g->textsize) {
   1198 				i /= HISTFRACTION * sizeof(*g->kcount);
   1199 				g->kcount[i]++;
   1200 			}
   1201 		}
   1202 #endif
   1203 #ifdef LWP_PC
   1204 		if (p && profsrc == PROFSRC_CLOCK && (p->p_flag & P_PROFIL))
   1205 			addupc_intr(p, LWP_PC(l));
   1206 #endif
   1207 		if (--spc->spc_pscnt > 0)
   1208 			return;
   1209 		/*
   1210 		 * Came from kernel mode, so we were:
   1211 		 * - handling an interrupt,
   1212 		 * - doing syscall or trap work on behalf of the current
   1213 		 *   user process, or
   1214 		 * - spinning in the idle loop.
   1215 		 * Whichever it is, charge the time as appropriate.
   1216 		 * Note that we charge interrupts to the current process,
   1217 		 * regardless of whether they are ``for'' that process,
   1218 		 * so that we know how much of its real time was spent
   1219 		 * in ``non-process'' (i.e., interrupt) work.
   1220 		 */
   1221 		if (CLKF_INTR(frame)) {
   1222 			if (p != NULL)
   1223 				p->p_iticks++;
   1224 			spc->spc_cp_time[CP_INTR]++;
   1225 		} else if (p != NULL) {
   1226 			p->p_sticks++;
   1227 			spc->spc_cp_time[CP_SYS]++;
   1228 		} else
   1229 			spc->spc_cp_time[CP_IDLE]++;
   1230 	}
   1231 	spc->spc_pscnt = psdiv;
   1232 
   1233 	if (p != NULL) {
   1234 		++p->p_cpticks;
   1235 		/*
   1236 		 * If no separate schedclock is provided, call it here
   1237 		 * at about 16 Hz.
   1238 		 */
   1239 		if (schedhz == 0)
   1240 			if ((int)(--ci->ci_schedstate.spc_schedticks) <= 0) {
   1241 				schedclock(l);
   1242 				ci->ci_schedstate.spc_schedticks = statscheddiv;
   1243 			}
   1244 	}
   1245 }
   1246 
   1247 #ifndef __HAVE_TIMECOUNTER
   1248 #ifdef NTP	/* NTP phase-locked loop in kernel */
   1249 /*
   1250  * hardupdate() - local clock update
   1251  *
   1252  * This routine is called by ntp_adjtime() to update the local clock
   1253  * phase and frequency. The implementation is of an adaptive-parameter,
   1254  * hybrid phase/frequency-lock loop (PLL/FLL). The routine computes new
   1255  * time and frequency offset estimates for each call. If the kernel PPS
   1256  * discipline code is configured (PPS_SYNC), the PPS signal itself
   1257  * determines the new time offset, instead of the calling argument.
   1258  * Presumably, calls to ntp_adjtime() occur only when the caller
   1259  * believes the local clock is valid within some bound (+-128 ms with
   1260  * NTP). If the caller's time is far different than the PPS time, an
   1261  * argument will ensue, and it's not clear who will lose.
   1262  *
   1263  * For uncompensated quartz crystal oscillatores and nominal update
   1264  * intervals less than 1024 s, operation should be in phase-lock mode
   1265  * (STA_FLL = 0), where the loop is disciplined to phase. For update
   1266  * intervals greater than thiss, operation should be in frequency-lock
   1267  * mode (STA_FLL = 1), where the loop is disciplined to frequency.
   1268  *
   1269  * Note: splclock() is in effect.
   1270  */
   1271 void
   1272 hardupdate(long offset)
   1273 {
   1274 	long ltemp, mtemp;
   1275 
   1276 	if (!(time_status & STA_PLL) && !(time_status & STA_PPSTIME))
   1277 		return;
   1278 	ltemp = offset;
   1279 #ifdef PPS_SYNC
   1280 	if (time_status & STA_PPSTIME && time_status & STA_PPSSIGNAL)
   1281 		ltemp = pps_offset;
   1282 #endif /* PPS_SYNC */
   1283 
   1284 	/*
   1285 	 * Scale the phase adjustment and clamp to the operating range.
   1286 	 */
   1287 	if (ltemp > MAXPHASE)
   1288 		time_offset = MAXPHASE << SHIFT_UPDATE;
   1289 	else if (ltemp < -MAXPHASE)
   1290 		time_offset = -(MAXPHASE << SHIFT_UPDATE);
   1291 	else
   1292 		time_offset = ltemp << SHIFT_UPDATE;
   1293 
   1294 	/*
   1295 	 * Select whether the frequency is to be controlled and in which
   1296 	 * mode (PLL or FLL). Clamp to the operating range. Ugly
   1297 	 * multiply/divide should be replaced someday.
   1298 	 */
   1299 	if (time_status & STA_FREQHOLD || time_reftime == 0)
   1300 		time_reftime = time.tv_sec;
   1301 	mtemp = time.tv_sec - time_reftime;
   1302 	time_reftime = time.tv_sec;
   1303 	if (time_status & STA_FLL) {
   1304 		if (mtemp >= MINSEC) {
   1305 			ltemp = ((time_offset / mtemp) << (SHIFT_USEC -
   1306 			    SHIFT_UPDATE));
   1307 			if (ltemp < 0)
   1308 				time_freq -= -ltemp >> SHIFT_KH;
   1309 			else
   1310 				time_freq += ltemp >> SHIFT_KH;
   1311 		}
   1312 	} else {
   1313 		if (mtemp < MAXSEC) {
   1314 			ltemp *= mtemp;
   1315 			if (ltemp < 0)
   1316 				time_freq -= -ltemp >> (time_constant +
   1317 				    time_constant + SHIFT_KF -
   1318 				    SHIFT_USEC);
   1319 			else
   1320 				time_freq += ltemp >> (time_constant +
   1321 				    time_constant + SHIFT_KF -
   1322 				    SHIFT_USEC);
   1323 		}
   1324 	}
   1325 	if (time_freq > time_tolerance)
   1326 		time_freq = time_tolerance;
   1327 	else if (time_freq < -time_tolerance)
   1328 		time_freq = -time_tolerance;
   1329 }
   1330 
   1331 #ifdef PPS_SYNC
   1332 /*
   1333  * hardpps() - discipline CPU clock oscillator to external PPS signal
   1334  *
   1335  * This routine is called at each PPS interrupt in order to discipline
   1336  * the CPU clock oscillator to the PPS signal. It measures the PPS phase
   1337  * and leaves it in a handy spot for the hardclock() routine. It
   1338  * integrates successive PPS phase differences and calculates the
   1339  * frequency offset. This is used in hardclock() to discipline the CPU
   1340  * clock oscillator so that intrinsic frequency error is cancelled out.
   1341  * The code requires the caller to capture the time and hardware counter
   1342  * value at the on-time PPS signal transition.
   1343  *
   1344  * Note that, on some Unix systems, this routine runs at an interrupt
   1345  * priority level higher than the timer interrupt routine hardclock().
   1346  * Therefore, the variables used are distinct from the hardclock()
   1347  * variables, except for certain exceptions: The PPS frequency pps_freq
   1348  * and phase pps_offset variables are determined by this routine and
   1349  * updated atomically. The time_tolerance variable can be considered a
   1350  * constant, since it is infrequently changed, and then only when the
   1351  * PPS signal is disabled. The watchdog counter pps_valid is updated
   1352  * once per second by hardclock() and is atomically cleared in this
   1353  * routine.
   1354  */
   1355 void
   1356 hardpps(struct timeval *tvp,		/* time at PPS */
   1357 	long usec			/* hardware counter at PPS */)
   1358 {
   1359 	long u_usec, v_usec, bigtick;
   1360 	long cal_sec, cal_usec;
   1361 
   1362 	/*
   1363 	 * An occasional glitch can be produced when the PPS interrupt
   1364 	 * occurs in the hardclock() routine before the time variable is
   1365 	 * updated. Here the offset is discarded when the difference
   1366 	 * between it and the last one is greater than tick/2, but not
   1367 	 * if the interval since the first discard exceeds 30 s.
   1368 	 */
   1369 	time_status |= STA_PPSSIGNAL;
   1370 	time_status &= ~(STA_PPSJITTER | STA_PPSWANDER | STA_PPSERROR);
   1371 	pps_valid = 0;
   1372 	u_usec = -tvp->tv_usec;
   1373 	if (u_usec < -500000)
   1374 		u_usec += 1000000;
   1375 	v_usec = pps_offset - u_usec;
   1376 	if (v_usec < 0)
   1377 		v_usec = -v_usec;
   1378 	if (v_usec > (tick >> 1)) {
   1379 		if (pps_glitch > MAXGLITCH) {
   1380 			pps_glitch = 0;
   1381 			pps_tf[2] = u_usec;
   1382 			pps_tf[1] = u_usec;
   1383 		} else {
   1384 			pps_glitch++;
   1385 			u_usec = pps_offset;
   1386 		}
   1387 	} else
   1388 		pps_glitch = 0;
   1389 
   1390 	/*
   1391 	 * A three-stage median filter is used to help deglitch the pps
   1392 	 * time. The median sample becomes the time offset estimate; the
   1393 	 * difference between the other two samples becomes the time
   1394 	 * dispersion (jitter) estimate.
   1395 	 */
   1396 	pps_tf[2] = pps_tf[1];
   1397 	pps_tf[1] = pps_tf[0];
   1398 	pps_tf[0] = u_usec;
   1399 	if (pps_tf[0] > pps_tf[1]) {
   1400 		if (pps_tf[1] > pps_tf[2]) {
   1401 			pps_offset = pps_tf[1];		/* 0 1 2 */
   1402 			v_usec = pps_tf[0] - pps_tf[2];
   1403 		} else if (pps_tf[2] > pps_tf[0]) {
   1404 			pps_offset = pps_tf[0];		/* 2 0 1 */
   1405 			v_usec = pps_tf[2] - pps_tf[1];
   1406 		} else {
   1407 			pps_offset = pps_tf[2];		/* 0 2 1 */
   1408 			v_usec = pps_tf[0] - pps_tf[1];
   1409 		}
   1410 	} else {
   1411 		if (pps_tf[1] < pps_tf[2]) {
   1412 			pps_offset = pps_tf[1];		/* 2 1 0 */
   1413 			v_usec = pps_tf[2] - pps_tf[0];
   1414 		} else  if (pps_tf[2] < pps_tf[0]) {
   1415 			pps_offset = pps_tf[0];		/* 1 0 2 */
   1416 			v_usec = pps_tf[1] - pps_tf[2];
   1417 		} else {
   1418 			pps_offset = pps_tf[2];		/* 1 2 0 */
   1419 			v_usec = pps_tf[1] - pps_tf[0];
   1420 		}
   1421 	}
   1422 	if (v_usec > MAXTIME)
   1423 		pps_jitcnt++;
   1424 	v_usec = (v_usec << PPS_AVG) - pps_jitter;
   1425 	if (v_usec < 0)
   1426 		pps_jitter -= -v_usec >> PPS_AVG;
   1427 	else
   1428 		pps_jitter += v_usec >> PPS_AVG;
   1429 	if (pps_jitter > (MAXTIME >> 1))
   1430 		time_status |= STA_PPSJITTER;
   1431 
   1432 	/*
   1433 	 * During the calibration interval adjust the starting time when
   1434 	 * the tick overflows. At the end of the interval compute the
   1435 	 * duration of the interval and the difference of the hardware
   1436 	 * counters at the beginning and end of the interval. This code
   1437 	 * is deliciously complicated by the fact valid differences may
   1438 	 * exceed the value of tick when using long calibration
   1439 	 * intervals and small ticks. Note that the counter can be
   1440 	 * greater than tick if caught at just the wrong instant, but
   1441 	 * the values returned and used here are correct.
   1442 	 */
   1443 	bigtick = (long)tick << SHIFT_USEC;
   1444 	pps_usec -= pps_freq;
   1445 	if (pps_usec >= bigtick)
   1446 		pps_usec -= bigtick;
   1447 	if (pps_usec < 0)
   1448 		pps_usec += bigtick;
   1449 	pps_time.tv_sec++;
   1450 	pps_count++;
   1451 	if (pps_count < (1 << pps_shift))
   1452 		return;
   1453 	pps_count = 0;
   1454 	pps_calcnt++;
   1455 	u_usec = usec << SHIFT_USEC;
   1456 	v_usec = pps_usec - u_usec;
   1457 	if (v_usec >= bigtick >> 1)
   1458 		v_usec -= bigtick;
   1459 	if (v_usec < -(bigtick >> 1))
   1460 		v_usec += bigtick;
   1461 	if (v_usec < 0)
   1462 		v_usec = -(-v_usec >> pps_shift);
   1463 	else
   1464 		v_usec = v_usec >> pps_shift;
   1465 	pps_usec = u_usec;
   1466 	cal_sec = tvp->tv_sec;
   1467 	cal_usec = tvp->tv_usec;
   1468 	cal_sec -= pps_time.tv_sec;
   1469 	cal_usec -= pps_time.tv_usec;
   1470 	if (cal_usec < 0) {
   1471 		cal_usec += 1000000;
   1472 		cal_sec--;
   1473 	}
   1474 	pps_time = *tvp;
   1475 
   1476 	/*
   1477 	 * Check for lost interrupts, noise, excessive jitter and
   1478 	 * excessive frequency error. The number of timer ticks during
   1479 	 * the interval may vary +-1 tick. Add to this a margin of one
   1480 	 * tick for the PPS signal jitter and maximum frequency
   1481 	 * deviation. If the limits are exceeded, the calibration
   1482 	 * interval is reset to the minimum and we start over.
   1483 	 */
   1484 	u_usec = (long)tick << 1;
   1485 	if (!((cal_sec == -1 && cal_usec > (1000000 - u_usec))
   1486 	    || (cal_sec == 0 && cal_usec < u_usec))
   1487 	    || v_usec > time_tolerance || v_usec < -time_tolerance) {
   1488 		pps_errcnt++;
   1489 		pps_shift = PPS_SHIFT;
   1490 		pps_intcnt = 0;
   1491 		time_status |= STA_PPSERROR;
   1492 		return;
   1493 	}
   1494 
   1495 	/*
   1496 	 * A three-stage median filter is used to help deglitch the pps
   1497 	 * frequency. The median sample becomes the frequency offset
   1498 	 * estimate; the difference between the other two samples
   1499 	 * becomes the frequency dispersion (stability) estimate.
   1500 	 */
   1501 	pps_ff[2] = pps_ff[1];
   1502 	pps_ff[1] = pps_ff[0];
   1503 	pps_ff[0] = v_usec;
   1504 	if (pps_ff[0] > pps_ff[1]) {
   1505 		if (pps_ff[1] > pps_ff[2]) {
   1506 			u_usec = pps_ff[1];		/* 0 1 2 */
   1507 			v_usec = pps_ff[0] - pps_ff[2];
   1508 		} else if (pps_ff[2] > pps_ff[0]) {
   1509 			u_usec = pps_ff[0];		/* 2 0 1 */
   1510 			v_usec = pps_ff[2] - pps_ff[1];
   1511 		} else {
   1512 			u_usec = pps_ff[2];		/* 0 2 1 */
   1513 			v_usec = pps_ff[0] - pps_ff[1];
   1514 		}
   1515 	} else {
   1516 		if (pps_ff[1] < pps_ff[2]) {
   1517 			u_usec = pps_ff[1];		/* 2 1 0 */
   1518 			v_usec = pps_ff[2] - pps_ff[0];
   1519 		} else  if (pps_ff[2] < pps_ff[0]) {
   1520 			u_usec = pps_ff[0];		/* 1 0 2 */
   1521 			v_usec = pps_ff[1] - pps_ff[2];
   1522 		} else {
   1523 			u_usec = pps_ff[2];		/* 1 2 0 */
   1524 			v_usec = pps_ff[1] - pps_ff[0];
   1525 		}
   1526 	}
   1527 
   1528 	/*
   1529 	 * Here the frequency dispersion (stability) is updated. If it
   1530 	 * is less than one-fourth the maximum (MAXFREQ), the frequency
   1531 	 * offset is updated as well, but clamped to the tolerance. It
   1532 	 * will be processed later by the hardclock() routine.
   1533 	 */
   1534 	v_usec = (v_usec >> 1) - pps_stabil;
   1535 	if (v_usec < 0)
   1536 		pps_stabil -= -v_usec >> PPS_AVG;
   1537 	else
   1538 		pps_stabil += v_usec >> PPS_AVG;
   1539 	if (pps_stabil > MAXFREQ >> 2) {
   1540 		pps_stbcnt++;
   1541 		time_status |= STA_PPSWANDER;
   1542 		return;
   1543 	}
   1544 	if (time_status & STA_PPSFREQ) {
   1545 		if (u_usec < 0) {
   1546 			pps_freq -= -u_usec >> PPS_AVG;
   1547 			if (pps_freq < -time_tolerance)
   1548 				pps_freq = -time_tolerance;
   1549 			u_usec = -u_usec;
   1550 		} else {
   1551 			pps_freq += u_usec >> PPS_AVG;
   1552 			if (pps_freq > time_tolerance)
   1553 				pps_freq = time_tolerance;
   1554 		}
   1555 	}
   1556 
   1557 	/*
   1558 	 * Here the calibration interval is adjusted. If the maximum
   1559 	 * time difference is greater than tick / 4, reduce the interval
   1560 	 * by half. If this is not the case for four consecutive
   1561 	 * intervals, double the interval.
   1562 	 */
   1563 	if (u_usec << pps_shift > bigtick >> 2) {
   1564 		pps_intcnt = 0;
   1565 		if (pps_shift > PPS_SHIFT)
   1566 			pps_shift--;
   1567 	} else if (pps_intcnt >= 4) {
   1568 		pps_intcnt = 0;
   1569 		if (pps_shift < PPS_SHIFTMAX)
   1570 			pps_shift++;
   1571 	} else
   1572 		pps_intcnt++;
   1573 }
   1574 #endif /* PPS_SYNC */
   1575 #endif /* NTP  */
   1576 
   1577 /* timecounter compat functions */
   1578 void
   1579 nanotime(struct timespec *ts)
   1580 {
   1581 	struct timeval tv;
   1582 
   1583 	microtime(&tv);
   1584 	TIMEVAL_TO_TIMESPEC(&tv, ts);
   1585 }
   1586 
   1587 void
   1588 getbinuptime(struct bintime *bt)
   1589 {
   1590 	struct timeval tv;
   1591 
   1592 	microtime(&tv);
   1593 	timeval2bintime(&tv, bt);
   1594 }
   1595 
   1596 void
   1597 nanouptime(struct timespec *tsp)
   1598 {
   1599 	int s;
   1600 
   1601 	s = splclock();
   1602 	TIMEVAL_TO_TIMESPEC(&mono_time, tsp);
   1603 	splx(s);
   1604 }
   1605 
   1606 void
   1607 getnanouptime(struct timespec *tsp)
   1608 {
   1609 	int s;
   1610 
   1611 	s = splclock();
   1612 	TIMEVAL_TO_TIMESPEC(&mono_time, tsp);
   1613 	splx(s);
   1614 }
   1615 
   1616 void
   1617 getmicrouptime(struct timeval *tvp)
   1618 {
   1619 	int s;
   1620 
   1621 	s = splclock();
   1622 	*tvp = mono_time;
   1623 	splx(s);
   1624 }
   1625 
   1626 void
   1627 getnanotime(struct timespec *tsp)
   1628 {
   1629 	int s;
   1630 
   1631 	s = splclock();
   1632 	TIMEVAL_TO_TIMESPEC(&time, tsp);
   1633 	splx(s);
   1634 }
   1635 
   1636 void
   1637 getmicrotime(struct timeval *tvp)
   1638 {
   1639 	int s;
   1640 
   1641 	s = splclock();
   1642 	*tvp = time;
   1643 	splx(s);
   1644 }
   1645 #endif /* !__HAVE_TIMECOUNTER */
   1646